In Gee (Chapter 7: The Social Mind in What Video Games Have to Teach us About Learning and Literacy) he extends the learning through video-games from single-player to multi-player.  As ever, he ends the chapter with a subset of the 36 Learning Principles (33-36) and these focus on the distributed nature of knowledge and the dispersal of this across 'affinity groups' - objects and systems as well as people - whereby he extends the notion of communities of practice and identity; all draw on traditional methods of storage and collation but importantly email, the internet, etc.

As with the classroom, players/participants work towards a common purpose, irrespective of race, class, gender, etc. This idea of common purpose and objective of all school pupils must be conveyed to pupils and allow them to acknowledge that it is their interest to work in partnership in order to support each other through their strengths and weaknesses (known as 'jigsaw method' defined by Brown & Campione (in Kate McQilly Ed. Classroom Lesson's: Integrating Cognitive Theory and Classroom Practice).

The idea of identity resurfaces, recalling the notion of 'multiple identity' whereby in such affinity groups "the norms and values of groups are contested and negotiated" (p. 184), bearing in mind that players - and us - modulate across a continuum of identity depending on context.  He finishes noting that this 'identity' should not only be in control but one who can modify content and the process itself.

These ideas accord with my periphery readings of Richard Riding's School Learning and Cognitive Styles where learners are negotiated into two dimensions of 'learning style' - wholist-analytic & visual-verbal.  In relation to the  affinity group that Gee promotes, Riding suggests that they, the learners, can work in partnerships that take into account differences / discrepancies in different learner types.  I think that this may have a significance in the successful presentation and reception of learning activities derived from digital games, and when you factor in collaborative learning I think that it is equally appropriate to look at compensatory pairing/grouping in addition.  

http://www.encorewiki.org/display/~hnajafi/Communities+of+Learners